Papers by Ruairí de Fréin
We apply Non-negative Matrix Factorization (NMF) to the problem of identifying underlying trends ... more We apply Non-negative Matrix Factorization (NMF) to the problem of identifying underlying trends in stock market data. NMF is a recent and very successful tool for data analysis including image and audio processing; we use it here to decompose a mixture a data, the daily closing prices of the 30 stocks which make up the Dow Jones Industrial Average, into its constitute parts, the underlying trends which Konstantinos Drakakis et al govern the financial marketplace. We demonstrate how to impose appropriate sparsity and smoothness constraints on the components of the decomposition. Also, we describe how the method clusters stocks together in performance-based groupings which can be used for portfolio diversification.
We apply Non-negative Matrix Factorization (NMF) to the prob- lem of identifying underlying trend... more We apply Non-negative Matrix Factorization (NMF) to the prob- lem of identifying underlying trends in stock market data. NMF is a recent and very successful tool for data analysis including image and audio processing; we use it here to decompose a mixture a data, the daily closing prices of the 30 stocks which make up the Dow Jones In- dustrial
We describe a new localisation and source separation algorithm which is based upon the accurate c... more We describe a new localisation and source separation algorithm which is based upon the accurate construction of time-frequency spatial signatures. We present a technique for constructing time-frequency spatial signatures with the required accuracy. This algorithm for multichannel source separation and localisation allows arbitrary placement of microphones yet achieves good performance. We demonstrate the efficacy of the technique using source location estimates and compare estimated time-frequency masks with the ideal 0 dB mask.
International Workshop on Machine Learning for Signal Processing
A central concern for many learning algorithms is how to efficiently store what the algorithm has... more A central concern for many learning algorithms is how to efficiently store what the algorithm has learned. An algorithm for the compression of Nonnegative Matrix Factorizations is presented. Compression is achieved by embedding the factorization in an encoding routine. Its performance is investigated using two standard test images, Peppers and Barbara. The compression ratio (18:1) achieved by the proposed Matrix Factorization improves the storage-ability of Nonnegative Matrix Factorizations without significantly degrading accuracy (. 1-3dB degradation is introduced). We learn as before, but storage is cheaper.
@INPROCEEDINGS{deFrein14Learning,
author={\href{http://robustandscalable.wordpress.com}{Ruair\'{i} {de Fr\'{e}in}}},
booktitle={IEEE International Workshop on Machine Learning for Signal Processing},
title={{Learning and Storing the Parts of Objects: IMF}},
year={2014},
month={Sep.},
pages={},
keywords={Approximation methods;Convergence;Digital signal processing;Matrix decomposition;Optimization;Quantization (signal);Signal processing algorithms;low rank;nmf;quantization},
doi={10.1109/ICDSP.2014.6900690},
}
Digital Signal Processing (DSP), 2014 19th International Conference on, Aug 2014
Even though Nonnegative Matrix Factorization (NMF) in its original form performs rank reduction a... more Even though Nonnegative Matrix Factorization (NMF) in its original form performs rank reduction and signal compaction implicitly, it does not explicitly consider storage or transmission constraints. We propose a Frobenius-norm Quantized Nonnegative Matrix Factorization algorithm that is 1) almost as precise as traditional NMF for decomposition ranks of interest (with in 1-4dB), 2) admits to practical encoding techniques by learning a factorization which is simpler than NMF's (by a factor of 20-70) and 3) exhibits a complexity which is comparable with state-of-the-art NMF methods. These properties are achieved by considering the quantization residual via an outer quantization optimization step, in an extended NMF iteration, namely QNMF. This approach comes in two forms: QNMF with 1) quasi-fixed and 2) adaptive quantization levels. Quantized NMF considers element-wise quantization constraints in the learning algorithm to eliminate defects due to post factorization quantization. We demonstrate significant reduction in the cardinality of the factor signal values set for comparable Signal-to-Noise-Ratios in a matrix decomposition task.
@INPROCEEDINGS{deFrein14Quantized,
author={\href{http://robustandscalable.wordpress.com}{Ruair\'{i} {de Fr\'{e}in}}},
booktitle={Digital Signal Processing (DSP), 2014 19th International Conference on},
title={{Quantized Nonnegative Matrix Factorization}},
year={2014},
month={Aug.},
pages={377--82},
keywords={Approximation methods;Convergence;Digital signal processing;Matrix decomposition;Optimization;Quantization (signal);Signal processing algorithms;low rank;nmf;quantization},
doi={10.1109/ICDSP.2014.6900690},
}
Journal of Network and Systems Management, 2014
Increasing and variable traffic demands due to triple play services pose significant Internet Pro... more Increasing and variable traffic demands due to triple play services pose significant Internet Protocol Television (IPTV) resource management challenges for service providers. Managing subscriber expectations via consolidated IPTV quality reporting will play a crucial role in guaranteeing return-on-investment for players in the increasingly competitive IPTV delivery ecosystem. We propose a fault diagnosis and problem isolation solution that addresses the IPTV monitoring challenge and recommends problem-specific remedial action. IPTV delivery-specific metrics are collected at various points in the delivery topology, the residential gateway and the Digital Subscriber Line Access Multiplexer through to the video Head-End. They are then pre-processed using new metric rules. A semantic uplift engine takes these raw metric logs; it then transforms them into World Wide Web Consortium’s standard Resource Description Framework for knowledge representation and annotates them with expert knowledge from the IPTV domain. This system is then integrated with a monitoring visualization framework that displays monitoring events, alarms, and recommends solutions. A suite of IPTV fault scenarios is presented and used to evaluate the feasibility of the solution. We demonstrate that professional service providers can provide timely reports on the quality of IPTV service delivery using this system.
@article{deFrein14Integration,
year={2014},
issn={1064-7570},
journal={Journal of Network and Systems Management},
doi={10.1007/s10922-014-9313-9},
title={{Integration of QoS Metrics, Rules and Semantic Uplift for Advanced IPTV Monitoring}},
url={http://dx.doi.org/10.1007/s10922-014-9313-9},
publisher={Springer US},
keywords={Network and system monitoring; Measures; Quality of Service; IPTV; Monitoring; Semantic uplift},
author={\href{http://robustandscalable.wordpress.com}{Ruair\'{i} {de Fr\'{e}in}} and Olariu, Cristian and Song, Yuqian and Brennan, Rob and McDonagh, Patrick and Hava, Adriana and Thorpe, Christina and Murphy, John and Murphy, Liam and French, Paul},
pages={1-36},
language={English}
}
Lecture Notes in Computer Science (Springer International Publishing, 2014), , Jun 2014
Formal Concept Analysis (FCA) decomposes a matrix into a set of sparse matrices capturing its und... more Formal Concept Analysis (FCA) decomposes a matrix into a set of sparse matrices capturing its underlying structure. A similar task for real-valued data, transform coding, arises in image compression. Existing cosine transform coding for JPEG image compression uses a fixed, decorrelating transform; however, compression is limited as images rarely consist of pure cosines. The question remains whether an FCA adaptive transform can be applied to image compression. We propose a multi-layer FCA (MFCA) adaptive ordered transform and Sequentially Sifted Linear Programming (SSLP) encoding pair for adaptive image compression. Our hypothesis is that MFCA’s sparse linear codes (closures) for natural scenes, are a complete family of ordered, localized, oriented, bandpass receptive fields, predicted by models of the primary visual cortex. Results on real data demonstrate that adaptive compression is feasible. These initial results may play a role in improving compression rates and extending the applicability of FCA to real-valued data.
@incollection{deFrein14Multilayer,
year={2014},
isbn={978-3-319-07247-0},
booktitle={Formal Concept Analysis},
volume={8478},
series={Lecture Notes in Computer Science},
editor={Glodeanu, Cynthia Vera and Kaytoue, Mehdi and Sacarea, Christian},
doi={10.1007/978-3-319-07248-7_18},
title={{Multilayered, Blocked Formal Concept Analyses for Adaptive Image Compression}},
url={http://dx.doi.org/10.1007/978-3-319-07248-7_18},
publisher={Springer International Publishing},
author={ \href{http://robustandscalable.wordpress.com}{Ruair\'{i} {de Fr\'{e}in}} },
pages={251--67},
language={English}
}
The 44th Annual IEEE/IFIP International Conference on Dependable Systems and Networks (DSN 2014), Georgia Tech, Atlanta, Georgia USA., Jun 2014
Biologically inspired networking approaches have attracted attention due to their simplicity and ... more Biologically inspired networking approaches have attracted attention due to their simplicity and efficiency as distributed problem solving methods. They are well suited to addressing challenges in existing and future computer networking scenarios. Scalability, adaptability and robustness, to name a few, are the advantageous properties common to many biologically inspired approaches. Once the appropriate biological mechanism has been identified, in many cases it can be used to produce efficient approaches to handle complexity, dynamic changes and failures in networks
Significant effort has been placed on analyzing and developing biologically inspired approaches and applying them to computer networking. This tutorial provides a comprehensive overview of biologically inspired approaches for the computer networking field. It emphasizes their primary characteristics as building blocks for designing distributed systems. First we characterize computer networking, its requirements, future perspectives and challenges. Next, we emphasize existing constraints in current networking algorithms and we highlight the advantages of bio-inspired approaches. Biologically inspired approaches for wide to nano computer networking are presented and categorized into “distributed system building blocks”. These building blocks include: communication, coordination and agreement, adaptation, robustness to attacks and failures, and synchronization. For each category/approach, we present the source of biological inspiration, the corresponding mathematical model, primary applications, advantages, limitations and potential directions for future applications. We conclude the tutorial with an overview of potential future directions for bio-inspired algorithms.
@misc {deFrein14Biologically,
title = {{Biologically Inspired Networking Approaches as Building Blocks for Distributed Systems.}},
year = "2014",
author = "Michele Nogueira and \href{http://robustandscalable.wordpress.com}{Ruair\'{i} de Fr\'{e}in} and Tonguz and Ozan",
howpublished = "The 44th Annual IEEE/IFIP International Conference on Dependable Systems and Networks (DSN 2014), Georgia Tech, Atlanta, Georgia USA.",
urllink = "http://2014.dsn.org/tutorials.shtml",
keywords = "Activator-inhibitor systems; Adaptability; Agreement; Artificial immune systems; Attacks and failures; Bio-inspired algorithms; Biological inspiration; Cellular signaling networks; Communication; Complexity; Coordination; Datataxis; Distributed system building blocks; Distributed Systems; Dynamic changes; Epidemic spreading; Failures; Faults; Firefly oscillators; Honey bee dance; Mathematical Model; Networking; Outages; Physiological networks; Quorum Sensing; Robust; Robustness; Scalability; Social insect routing; Synchronization."}
Signals and Systems Conference (ISSC 2013), 24th IET Irish, Jun 2013
An exact nonnegative matrix decomposition algorithm is proposed. This is achieved by 1) Taking a ... more An exact nonnegative matrix decomposition algorithm is proposed. This is achieved by 1) Taking a nonlinear approximation of a sparse real-valued dataset at a given tolerance-to-error constraint, ε; Choosing an arbitrary lectic ordering on the rows or column entries; And, then systematically applying a closure operator, so that all closures are selected. Assuming a nonnegative hierarchical closure structure (a Galois lattice) ensures the data has a ordered overcomplete dictionary representation. Parts-based constraints on these closures can then be used to specify and supervise the form of the solution. We illustrate that this approach outperforms NMF on two standard NMF datasets: it exhibits the properties described above; It is correct and exact.
@inproceedings {deFrein13Ghostbusters,
title = {{Ghostbusters: A parts-based NMF algorithm}},
year = "2013",
author = "\href{http://robustandscalable.wordpress.com}{Ruair\'{i} de Fr\'{e}in}",
booktitle = "Signals and Systems Conference (ISSC 2013), 24th IET Irish",
pages = "1-8",
organization = "The Institution of Engineering and Technology - The IET",
doi = "10.1049/ic.2013.0050",
url = "http://repository.wit.ie/2745/1/RdFISSC.pdf",
urllink = "http://repository.wit.ie/2745/1/RdFISSC.pdf",
keywords = "image coding;matrix decomposition;Galois lattice;Ghostbusters;arbitrary lectic ordering;closure operator;column entry;dictionary representation;encoding part-based constraints;nonlinear approximation;nonnegative hierarchical closure structure;nonnegative matrix decomposition algorithm;part-based NMF algorithm;row entry;sparse real-valued dataset;tolerance-to-error constraint;Lectic Orderings;Nonnegative Matrix Factorization;Unique Solutions",
month = "Jun."}
WO Patent , Jan 2013
@misc {deFrein13Method,
title = {{METHOD AND APPARATUS FOR DETERMINING WHETHER A NO... more @misc {deFrein13Method,
title = {{METHOD AND APPARATUS FOR DETERMINING WHETHER A NODE CAN REPRESENT OR BE REPRESENTED BY OTHER NODES WITHIN A NETWORK}},
year = "2013",
author = "\href{http://robustandscalable.wordpress.com}{Ruair\'{i} de Fr\'{e}in}",
howpublished = "Patent",
url = "http://patentscope.wipo.int/search/en/WO2013007521",
urllink = "http://patentscope.wipo.int/search/en/WO2013007521",
note = "WO Patent 2,013,007,521, Patent Number 2013007521",
month = "18 Jan."}
Formal Concept Analysis, Jun 2013
Formal Concept Analysis (FCA) looks to decompose a matrix of objects-attributes into a set of spa... more Formal Concept Analysis (FCA) looks to decompose a matrix of objects-attributes into a set of sparse matrices capturing the underlying structure of a formal context. We propose a Rank Reduction (RR) method to prime approximate FCAs, namely RRFCA. While many existing FCA algorithms are complete, lectic ordering of the lattice may not minimize search/decomposition time. Initially, RRFCA decompositions are not unique or complete; however, a set of good closures with high support is learned quickly, and then, made complete. RRFCA has its novelty in that we propose a new multiplicative two-stage method. First, we describe the theoretical foundations underpinning our RR approach. Second, we provide a representative exemplar, showing how RRFCA can be implemented. Further experiments demonstrate that RRFCA methods are efficient, scalable and yield time-savings. We demonstrate the resulting methods lend themselves to parallelisation.
@incollection {deFrein2013Formal,
title = {{Formal Concept Analysis via Atomic Priming}},
year = "2013",
author = "\href{http://robustandscalable.wordpress.com}{Ruair\'{i} de Fr\'{e}in}",
booktitle = "Formal Concept Analysis",
editor = "Peggy Cellier and Felix Distel and Bernhard Ganter",
pages = "92-108",
organization = "Lecture Notes in Computer Science",
publisher = "Springer Berlin Heidelberg",
doi = {10.1007/978-3-642-38317-5_6},
urllink = "http://dx.doi.org/10.1007/978-3-642-38317-5\_6",
keywords = "Formal Concept Analysis; Rank Reduction; Factorization",
volume = "7880",
howpublished = "http://dx.doi.org/10.1007/978-3-642-38317-5\_6"}
Formal Concept Analysis, 2012
While many existing formal concept analysis algorithms are efficient, they are typically unsuitab... more While many existing formal concept analysis algorithms are efficient, they are typically unsuitable for distributed implementation. Taking the MapReduce (MR) framework as our inspiration we introduce a distributed approach for performing formal concept mining. Our method has its novelty in that we use a light-weight MapReduce runtime called Twister which is better suited to iterative algorithms than recent distributed approaches. First, we describe the theoretical foundations underpinning our distributed formal concept analysis approach. Second, we provide a representative exemplar of how a classic centralized algorithm can be implemented in a distributed fashion using our methodology: we modify Ganter’s classic algorithm by introducing a family of MR∗ algorithms, namely MRGanter and MRGanter+ where the prefix denotes the algorithm’s lineage. To evaluate the factors that impact distributed algorithm performance, we compare our MR∗ algorithms with the state-of-the-art. Experiments conducted on real datasets demonstrate that MRGanter+ is efficient, scalable and an appealing algorithm for distributed problems.
@incollection {Xu2012Distributed,
title = {{Distributed Formal Concept Analysis Algorithms Based on an Iterative MapReduce Framework}},
year = "2012",
author = "Biao Xu and \href{http://robustandscalable.wordpress.com}{Ruair\'{i} de Fr\'{e}in} and Eric Robson and M\'{i}che\'{a}l \'{O} Foghl\'{u}",
booktitle = "Formal Concept Analysis",
editor = "Florent Domenach and DmitryI Ignatov and Jonas Poelmans",
pages = "292-308",
organization = "Lecture Notes in Computer Science, 7278, ",
publisher = "Springer Berlin Heidelberg",
doi = "10.1007/978-3-642-29892-9_26",
urllink = "http://dx.doi.org/10.1007/978-3-642-29892-9_26",
keywords = "Formal Concept Analysis; Distributed Mining; MapReduce",
volume = "7278",
howpublished = "http://dx.doi.org/10.1007/978-3-642-29892-9\_26"}
MRGanter+is a distributed Formal Concept Analysis algorithm based on Gater's algorithm (as known ... more MRGanter+is a distributed Formal Concept Analysis algorithm based on Gater's algorithm (as known as NextClosure) and an iterative MapReduce framework, Twister. NextClosure calculates closures in lectic ordering to ensure every concept appears only once. This approach allows a single concept to be tested with the closure validation condition during each iteration. This is efficient when the algorithm runs on a single machine. For multi-machine computation, the extra computation and redundancy resulting from keeping only one concept after each iteration across many machines is costly. We modify NextClosure to reduce the number of iterations and name the corresponding distributed algorithm MRGanter+.
Rather than using redundancy checking, we keep as many closures as possible in each iteration; All closures are maintained and used to generate the next batch of closures. MRGanter+ has a Map method which calculates local concepts by working on previous concept and local data partition. The Reduce method in MRGanter+ merges local closures first, and then recursively examines if they already exist in the set of global formal concepts H. The set H is used to fast index and search a specified closure; it is designed as a two-level hash table to reduce its costs. The first level is indexed by the head attribute of the closure, while the second level is indexed by the length of the closure.
For more details about MRGanter+, please see our recent publication:
Distributed Formal Concept Analysis Algorithms Based on an Iterative MapReduce Framework:
@incollection{,
year={2012},
isbn={978-3-642-29891-2},
booktitle={Formal Concept Analysis},
volume={7278},
series={Lecture Notes in Computer Science},
editor={Domenach, Florent and Ignatov, DmitryI. and Poelmans, Jonas},
doi={10.1007/978-3-642-29892-9_26},
title={Distributed Formal Concept Analysis Algorithms Based on an Iterative MapReduce Framework},
url={http://dx.doi.org/10.1007/978-3-642-29892-9_26},
publisher={Springer Berlin Heidelberg},
keywords={Formal Concept Analysis; Distributed Mining; MapReduce},
author={Xu, Biao and Fréin, Ruairí and Robson, Eric and Ó Foghlú, Mícheál},
pages={292-308}
}
at http://www.springerlink.com/content/02p8282703rx0m78/. And you can download the author's self-archiving version of this publication from http://arxiv.org/abs/1210.2401.
To know how MapReduce works please refer to "MapReduce Simplified Data Processing on Large Clusters" at http://static.usenix.org/event/osdi04/tech/full_papers/dean/dean.pdf.
Or you can find a presentaion introducing MRGanter+ which we made on ICFCA 2012: http://www.econ.kuleuven.be/ICFCA/present/Biao_Xu.pdf
IEEE Transactions on Signal Processing, 2011
This paper proposes the use of a synchronized linear transform, the synchronized short-time-Fouri... more This paper proposes the use of a synchronized linear transform, the synchronized short-time-Fourier-transform (sSTFT), for time-frequency analysis of anechoic mixtures. We address the short comings of the commonly used time-frequency linear transform in multichannel settings, namely the classical short-time-Fourier-transform (cSTFT). We propose a series of desirable properties for the linear transform used in a multichannel source separation scenario: stationary invertibility, relative delay, relative attenuation, and finally delay invariant relative windowed-disjoint orthogonality (DIRWDO). Multisensor source separation techniques which operate in the time-frequency domain, have an inherent error unless consideration is given to the multichannel properties proposed in this paper. The sSTFT preserves these relationships for multichannel data. The crucial innovation of the sSTFT is to locally synchronize the analysis to the observations as opposed to a global clock. Improvement in separation performance can be achieved because assumed properties of the time-frequency transform are satisfied when it is appropriately synchronized. Numerical experiments show the sSTFT improves instantaneous subsample relative parameter estimation in low noise conditions and achieves good synthesis.
@ARTICLE{deFrein11Synchronized,
author={\href{http://robustandscalable.wordpress.com}{Ruair\'{i} {de Fr\'{e}in}} and Rickard, Scott T.},
journal={Signal Processing, IEEE Transactions on},
title={{The Synchronized Short-Time-Fourier-Transform: Properties and Definitions for Multichannel Source Separation}},
year={2011},
volume={59},
number={1},
pages={91--103},
keywords={Fourier transforms;parameter estimation;source separation;synchronisation;anechoic mixtures;delay invariant relative windowed-disjoint orthogonality;multichannel source separation;parameter estimation;short time Fourier transform;synchronized linear transform;time frequency domain;time frequency linear transform;Delay;Fourier transforms;Signal processing algorithms;Speech;Synchronization;Time frequency analysis;Signal analysis;source separation},
doi={10.1109/TSP.2010.2088392},
ISSN={1053-587X},}
IEEE Transactions on Signal Processing, 2011
This paper proposes the use of a synchronized linear transform, the synchronized short-time-Fouri... more This paper proposes the use of a synchronized linear transform, the synchronized short-time-Fourier-transform (sSTFT), for time-frequency analysis of anechoic mixtures. We address the short comings of the commonly used time-frequency linear transform in multichannel settings, namely the classical short-time-Fourier-transform (cSTFT). We propose a series of desirable properties for the linear transform used in a multichannel source separation scenario: stationary invertibility, relative delay, relative attenuation, and finally delay invariant relative windowed-disjoint orthogonality (DIRWDO). Multisensor source separation techniques which operate in the time-frequency domain, have an inherent error unless consideration is given to the multichannel properties proposed in this paper. The sSTFT preserves these relationships for multichannel data. The crucial innovation of the sSTFT is to locally synchronize the analysis to the observations as opposed to a global clock. Improvement in separation performance can be achieved because assumed properties of the time-frequency transform are satisfied when it is appropriately synchronized. Numerical experiments show the sSTFT improves instantaneous subsample relative parameter estimation in low noise conditions and achieves good synthesis.
PhD Thesis, Dec 2010
@phdthesis {deFrein10Adapting,
title = {{Adapting Bases Using the Synchronized Shor... more @phdthesis {deFrein10Adapting,
title = {{Adapting Bases Using the Synchronized Short Time Fourier Transform and Non-negative Matrix Factorization}},
year = "2010",
author = "\href{http://robustandscalable.wordpress.com}{Ruair\'{i} de Fr\'{e}in}",
school = "University College Dublin"}
Digital Signal Processing, 2009 16th International Conference on, 2009
We introduce a non-negative matrix factorization technique which learns speech features with temp... more We introduce a non-negative matrix factorization technique which learns speech features with temporal extent in the presence of non-stationary noise. Our proposed technique, namely Sparse convolutive robust non-negative matrix factorization, is robust in the presence of noise due to our explicit treatment of noise as an interfering source in the factorization. We derive multiplicative update rules using the alpha divergence objective. We show that our proposed method yields superior performance to sparse convolutive non-negative matrix factorization in a feature learning task on noisy data and comparable results to dedicated speech enhancement techniques.
@inproceedings {deFrein09Learning,
title = {{Learning Speech Features in the Presence of Noise: Sparse Convolutive Robust Non-negative Matrix Factorisation}},
year = "2009",
author = "\href{http://robustandscalable.wordpress.com}{Ruair\'{i} de Fr\'{e}in} and Scott T. Rickard",
booktitle = "Digital Signal Processing, 2009 16th International Conference on",
pages = "1--6",
organization = "IEEE",
doi = "10.1109/ICDSP.2009.5201068",
urllink = "http://ieeexplore.ieee.org/xpl/login.jsp?tp=&arnumber=5201068&url=http%3A%2F%2Fieeexplore.ieee.org%2Fxpls%2Fabs_all.jsp%3Farnumber%3D5201068",
keywords = "matrix algebra;speech enhancement;alpha divergence objective;nonstationary noise;sparse convolutive robust non-negative matrix factorization;speech enhancement techniques;speech features;Additive noise;Background noise;Matrix decomposition;Noise reduction;Noise robustness;Sparse matrices;Spectrogram;Speech analysis;Speech enhancement;Working environment noise;Spectral factorization;Speech enhancement",
month = "Jul."}
Independent Component Analysis and Signal Separation, 2009
We describe a new localisation and source separation algorithm which is based upon the accurate c... more We describe a new localisation and source separation algorithm which is based upon the accurate construction of time-frequency spatial signatures. We present a technique for constructing time-frequency spatial signatures with the required accuracy. This algorithm for multi-channel source separation and localisation allows arbitrary placement of microphones yet achieves good performance. We demonstrate the efficacy of the technique using source location estimates and compare estimated time-frequency masks with the ideal 0 dB mask.
@inproceedings {deFrein09Constructing,
title = {{Constructing Time-Frequency Dictionaries for Source Separation via Time-Frequency Masking and Source Localisation}},
year = "2009",
author = "\href{http://robustandscalable.wordpress.com}{Ruair\'{i} de Fr\'{e}in} and Scott T. Rickard and Barak A. Pearlmutter",
booktitle = "Independent Component Analysis and Signal Separation",
editor = " Tulay Adali and Christian Jutten and Joao Marcos Romano and Travassos and Allan Kardec Barros",
pages = "573--80",
organization = "Lecture Notes in Computer Science, 7278,",
publisher = "Springer Berlin Heidelberg",
doi = "10.1007/978-3-642-00599-2_72",
url = "http://www.google.fr/url?sa=t&rct=j&q=&esrc=s&source=web&cd=2&ved=0CC8QFjAB&url=http%3A%2F%2Fwww.bcl.hamilton.ie%2F~barak%2Fpapers%2FICA-2009-constructing.pdf&ei=pZ1LVLasMomrOpTKgZgO&usg=AFQjCNHK8kTyiwRgq0jWwCU62yqyC0uyjg&bvm=bv.77880786,d.ZWU",
urllink = "http://dx.doi.org/10.1007/978-3-642-00599-2_72",
volume = "5441",
howpublished = "http://dx.doi.org/10.1007/978-3-642-00599-2_72"}
Digital Signal Processing Workshop and 5th IEEE Signal Processing Education Workshop, 2009
Identification of assets on the stock market that exhibit co-movement is a critical task for gene... more Identification of assets on the stock market that exhibit co-movement is a critical task for generating an efficiently diversified portfolio. We present a new application of non-negative matrix factorization to factor analysis of financial time series. We consider a conditionally heteroscedastic latent factor model, where each series is parameterized by a univariate ARCH model. Volatility clustering characteristics, e.g. GARCH effects, of the constituent assets of the Dow Jones Industrial Average are lever-aged to cluster assets based on the commonality of their volatility clusters. We present a new non-negative matrix factorization algorithm which is robust in the presence of noise, Robust NMF. We use a mixed low-rank over-complete dictionary learning approach to separate out the background Gaussian noise, emphasize the GARCH effects and achieve clearer asset groupings.
@inproceedings {deFrein09Extracting,
title = {{Extracting Garch Effects from Asset Returns Using Robust NMF}},
year = "2009",
author = "\href{http://robustandscalable.wordpress.com}{Ruair\'{i} de Fr\'{e}in} and Scott Rickard and Konstantinos Drakakis",
booktitle = "Digital Signal Processing Workshop and 5th IEEE Signal Processing Education Workshop, 2009. DSP/SPE 2009. IEEE 13th",
pages = "200--5",
organization = "IEEE",
doi = "10.1109/DSP.2009.4785921",
urllink = "http://ieeexplore.ieee.org/xpl/login.jsp?tp=&arnumber=4785921&url=http%3A%2F%2Fieeexplore.ieee.org%2Fxpls%2Fabs_all.jsp%3Farnumber%3D4785921",
keywords = "Gaussian noise;learning (artificial intelligence);matrix decomposition;pattern clustering;stock markets;time series;Dow Jones Industrial Average;GARCH effects;asset groupings;asset identification;asset returns;background Gaussian noise;cluster assets;dictionary learning approach;diversified portfolio;factor analysis;financial time series;heteroscedastic latent factor model;non-negative matrix factorization;robust NMF;stock market;univariate ARCH model;volatility clustering characteristics;Adaptive systems;Autocorrelation;Educational institutions;Laboratories;Matrix decomposition;Noise robustness;Portfolios;Probability distribution;Stock markets;Time series analysis;Autoregressive Conditional Heteroscedasticity;Clustering;Low rank decomposition;Non-negative Matrix Factorization;Sparseness",
month = "Jan."}
Uploads
Papers by Ruairí de Fréin
@INPROCEEDINGS{deFrein14Learning,
author={\href{http://robustandscalable.wordpress.com}{Ruair\'{i} {de Fr\'{e}in}}},
booktitle={IEEE International Workshop on Machine Learning for Signal Processing},
title={{Learning and Storing the Parts of Objects: IMF}},
year={2014},
month={Sep.},
pages={},
keywords={Approximation methods;Convergence;Digital signal processing;Matrix decomposition;Optimization;Quantization (signal);Signal processing algorithms;low rank;nmf;quantization},
doi={10.1109/ICDSP.2014.6900690},
}
@INPROCEEDINGS{deFrein14Quantized,
author={\href{http://robustandscalable.wordpress.com}{Ruair\'{i} {de Fr\'{e}in}}},
booktitle={Digital Signal Processing (DSP), 2014 19th International Conference on},
title={{Quantized Nonnegative Matrix Factorization}},
year={2014},
month={Aug.},
pages={377--82},
keywords={Approximation methods;Convergence;Digital signal processing;Matrix decomposition;Optimization;Quantization (signal);Signal processing algorithms;low rank;nmf;quantization},
doi={10.1109/ICDSP.2014.6900690},
}
@article{deFrein14Integration,
year={2014},
issn={1064-7570},
journal={Journal of Network and Systems Management},
doi={10.1007/s10922-014-9313-9},
title={{Integration of QoS Metrics, Rules and Semantic Uplift for Advanced IPTV Monitoring}},
url={http://dx.doi.org/10.1007/s10922-014-9313-9},
publisher={Springer US},
keywords={Network and system monitoring; Measures; Quality of Service; IPTV; Monitoring; Semantic uplift},
author={\href{http://robustandscalable.wordpress.com}{Ruair\'{i} {de Fr\'{e}in}} and Olariu, Cristian and Song, Yuqian and Brennan, Rob and McDonagh, Patrick and Hava, Adriana and Thorpe, Christina and Murphy, John and Murphy, Liam and French, Paul},
pages={1-36},
language={English}
}
@incollection{deFrein14Multilayer,
year={2014},
isbn={978-3-319-07247-0},
booktitle={Formal Concept Analysis},
volume={8478},
series={Lecture Notes in Computer Science},
editor={Glodeanu, Cynthia Vera and Kaytoue, Mehdi and Sacarea, Christian},
doi={10.1007/978-3-319-07248-7_18},
title={{Multilayered, Blocked Formal Concept Analyses for Adaptive Image Compression}},
url={http://dx.doi.org/10.1007/978-3-319-07248-7_18},
publisher={Springer International Publishing},
author={ \href{http://robustandscalable.wordpress.com}{Ruair\'{i} {de Fr\'{e}in}} },
pages={251--67},
language={English}
}
Significant effort has been placed on analyzing and developing biologically inspired approaches and applying them to computer networking. This tutorial provides a comprehensive overview of biologically inspired approaches for the computer networking field. It emphasizes their primary characteristics as building blocks for designing distributed systems. First we characterize computer networking, its requirements, future perspectives and challenges. Next, we emphasize existing constraints in current networking algorithms and we highlight the advantages of bio-inspired approaches. Biologically inspired approaches for wide to nano computer networking are presented and categorized into “distributed system building blocks”. These building blocks include: communication, coordination and agreement, adaptation, robustness to attacks and failures, and synchronization. For each category/approach, we present the source of biological inspiration, the corresponding mathematical model, primary applications, advantages, limitations and potential directions for future applications. We conclude the tutorial with an overview of potential future directions for bio-inspired algorithms.
@misc {deFrein14Biologically,
title = {{Biologically Inspired Networking Approaches as Building Blocks for Distributed Systems.}},
year = "2014",
author = "Michele Nogueira and \href{http://robustandscalable.wordpress.com}{Ruair\'{i} de Fr\'{e}in} and Tonguz and Ozan",
howpublished = "The 44th Annual IEEE/IFIP International Conference on Dependable Systems and Networks (DSN 2014), Georgia Tech, Atlanta, Georgia USA.",
urllink = "http://2014.dsn.org/tutorials.shtml",
keywords = "Activator-inhibitor systems; Adaptability; Agreement; Artificial immune systems; Attacks and failures; Bio-inspired algorithms; Biological inspiration; Cellular signaling networks; Communication; Complexity; Coordination; Datataxis; Distributed system building blocks; Distributed Systems; Dynamic changes; Epidemic spreading; Failures; Faults; Firefly oscillators; Honey bee dance; Mathematical Model; Networking; Outages; Physiological networks; Quorum Sensing; Robust; Robustness; Scalability; Social insect routing; Synchronization."}
@inproceedings {deFrein13Ghostbusters,
title = {{Ghostbusters: A parts-based NMF algorithm}},
year = "2013",
author = "\href{http://robustandscalable.wordpress.com}{Ruair\'{i} de Fr\'{e}in}",
booktitle = "Signals and Systems Conference (ISSC 2013), 24th IET Irish",
pages = "1-8",
organization = "The Institution of Engineering and Technology - The IET",
doi = "10.1049/ic.2013.0050",
url = "http://repository.wit.ie/2745/1/RdFISSC.pdf",
urllink = "http://repository.wit.ie/2745/1/RdFISSC.pdf",
keywords = "image coding;matrix decomposition;Galois lattice;Ghostbusters;arbitrary lectic ordering;closure operator;column entry;dictionary representation;encoding part-based constraints;nonlinear approximation;nonnegative hierarchical closure structure;nonnegative matrix decomposition algorithm;part-based NMF algorithm;row entry;sparse real-valued dataset;tolerance-to-error constraint;Lectic Orderings;Nonnegative Matrix Factorization;Unique Solutions",
month = "Jun."}
title = {{METHOD AND APPARATUS FOR DETERMINING WHETHER A NODE CAN REPRESENT OR BE REPRESENTED BY OTHER NODES WITHIN A NETWORK}},
year = "2013",
author = "\href{http://robustandscalable.wordpress.com}{Ruair\'{i} de Fr\'{e}in}",
howpublished = "Patent",
url = "http://patentscope.wipo.int/search/en/WO2013007521",
urllink = "http://patentscope.wipo.int/search/en/WO2013007521",
note = "WO Patent 2,013,007,521, Patent Number 2013007521",
month = "18 Jan."}
@incollection {deFrein2013Formal,
title = {{Formal Concept Analysis via Atomic Priming}},
year = "2013",
author = "\href{http://robustandscalable.wordpress.com}{Ruair\'{i} de Fr\'{e}in}",
booktitle = "Formal Concept Analysis",
editor = "Peggy Cellier and Felix Distel and Bernhard Ganter",
pages = "92-108",
organization = "Lecture Notes in Computer Science",
publisher = "Springer Berlin Heidelberg",
doi = {10.1007/978-3-642-38317-5_6},
urllink = "http://dx.doi.org/10.1007/978-3-642-38317-5\_6",
keywords = "Formal Concept Analysis; Rank Reduction; Factorization",
volume = "7880",
howpublished = "http://dx.doi.org/10.1007/978-3-642-38317-5\_6"}
@incollection {Xu2012Distributed,
title = {{Distributed Formal Concept Analysis Algorithms Based on an Iterative MapReduce Framework}},
year = "2012",
author = "Biao Xu and \href{http://robustandscalable.wordpress.com}{Ruair\'{i} de Fr\'{e}in} and Eric Robson and M\'{i}che\'{a}l \'{O} Foghl\'{u}",
booktitle = "Formal Concept Analysis",
editor = "Florent Domenach and DmitryI Ignatov and Jonas Poelmans",
pages = "292-308",
organization = "Lecture Notes in Computer Science, 7278, ",
publisher = "Springer Berlin Heidelberg",
doi = "10.1007/978-3-642-29892-9_26",
urllink = "http://dx.doi.org/10.1007/978-3-642-29892-9_26",
keywords = "Formal Concept Analysis; Distributed Mining; MapReduce",
volume = "7278",
howpublished = "http://dx.doi.org/10.1007/978-3-642-29892-9\_26"}
Rather than using redundancy checking, we keep as many closures as possible in each iteration; All closures are maintained and used to generate the next batch of closures. MRGanter+ has a Map method which calculates local concepts by working on previous concept and local data partition. The Reduce method in MRGanter+ merges local closures first, and then recursively examines if they already exist in the set of global formal concepts H. The set H is used to fast index and search a specified closure; it is designed as a two-level hash table to reduce its costs. The first level is indexed by the head attribute of the closure, while the second level is indexed by the length of the closure.
For more details about MRGanter+, please see our recent publication:
Distributed Formal Concept Analysis Algorithms Based on an Iterative MapReduce Framework:
@incollection{,
year={2012},
isbn={978-3-642-29891-2},
booktitle={Formal Concept Analysis},
volume={7278},
series={Lecture Notes in Computer Science},
editor={Domenach, Florent and Ignatov, DmitryI. and Poelmans, Jonas},
doi={10.1007/978-3-642-29892-9_26},
title={Distributed Formal Concept Analysis Algorithms Based on an Iterative MapReduce Framework},
url={http://dx.doi.org/10.1007/978-3-642-29892-9_26},
publisher={Springer Berlin Heidelberg},
keywords={Formal Concept Analysis; Distributed Mining; MapReduce},
author={Xu, Biao and Fréin, Ruairí and Robson, Eric and Ó Foghlú, Mícheál},
pages={292-308}
}
at http://www.springerlink.com/content/02p8282703rx0m78/. And you can download the author's self-archiving version of this publication from http://arxiv.org/abs/1210.2401.
To know how MapReduce works please refer to "MapReduce Simplified Data Processing on Large Clusters" at http://static.usenix.org/event/osdi04/tech/full_papers/dean/dean.pdf.
Or you can find a presentaion introducing MRGanter+ which we made on ICFCA 2012: http://www.econ.kuleuven.be/ICFCA/present/Biao_Xu.pdf
@ARTICLE{deFrein11Synchronized,
author={\href{http://robustandscalable.wordpress.com}{Ruair\'{i} {de Fr\'{e}in}} and Rickard, Scott T.},
journal={Signal Processing, IEEE Transactions on},
title={{The Synchronized Short-Time-Fourier-Transform: Properties and Definitions for Multichannel Source Separation}},
year={2011},
volume={59},
number={1},
pages={91--103},
keywords={Fourier transforms;parameter estimation;source separation;synchronisation;anechoic mixtures;delay invariant relative windowed-disjoint orthogonality;multichannel source separation;parameter estimation;short time Fourier transform;synchronized linear transform;time frequency domain;time frequency linear transform;Delay;Fourier transforms;Signal processing algorithms;Speech;Synchronization;Time frequency analysis;Signal analysis;source separation},
doi={10.1109/TSP.2010.2088392},
ISSN={1053-587X},}
title = {{Adapting Bases Using the Synchronized Short Time Fourier Transform and Non-negative Matrix Factorization}},
year = "2010",
author = "\href{http://robustandscalable.wordpress.com}{Ruair\'{i} de Fr\'{e}in}",
school = "University College Dublin"}
@inproceedings {deFrein09Learning,
title = {{Learning Speech Features in the Presence of Noise: Sparse Convolutive Robust Non-negative Matrix Factorisation}},
year = "2009",
author = "\href{http://robustandscalable.wordpress.com}{Ruair\'{i} de Fr\'{e}in} and Scott T. Rickard",
booktitle = "Digital Signal Processing, 2009 16th International Conference on",
pages = "1--6",
organization = "IEEE",
doi = "10.1109/ICDSP.2009.5201068",
urllink = "http://ieeexplore.ieee.org/xpl/login.jsp?tp=&arnumber=5201068&url=http%3A%2F%2Fieeexplore.ieee.org%2Fxpls%2Fabs_all.jsp%3Farnumber%3D5201068",
keywords = "matrix algebra;speech enhancement;alpha divergence objective;nonstationary noise;sparse convolutive robust non-negative matrix factorization;speech enhancement techniques;speech features;Additive noise;Background noise;Matrix decomposition;Noise reduction;Noise robustness;Sparse matrices;Spectrogram;Speech analysis;Speech enhancement;Working environment noise;Spectral factorization;Speech enhancement",
month = "Jul."}
@inproceedings {deFrein09Constructing,
title = {{Constructing Time-Frequency Dictionaries for Source Separation via Time-Frequency Masking and Source Localisation}},
year = "2009",
author = "\href{http://robustandscalable.wordpress.com}{Ruair\'{i} de Fr\'{e}in} and Scott T. Rickard and Barak A. Pearlmutter",
booktitle = "Independent Component Analysis and Signal Separation",
editor = " Tulay Adali and Christian Jutten and Joao Marcos Romano and Travassos and Allan Kardec Barros",
pages = "573--80",
organization = "Lecture Notes in Computer Science, 7278,",
publisher = "Springer Berlin Heidelberg",
doi = "10.1007/978-3-642-00599-2_72",
url = "http://www.google.fr/url?sa=t&rct=j&q=&esrc=s&source=web&cd=2&ved=0CC8QFjAB&url=http%3A%2F%2Fwww.bcl.hamilton.ie%2F~barak%2Fpapers%2FICA-2009-constructing.pdf&ei=pZ1LVLasMomrOpTKgZgO&usg=AFQjCNHK8kTyiwRgq0jWwCU62yqyC0uyjg&bvm=bv.77880786,d.ZWU",
urllink = "http://dx.doi.org/10.1007/978-3-642-00599-2_72",
volume = "5441",
howpublished = "http://dx.doi.org/10.1007/978-3-642-00599-2_72"}
@inproceedings {deFrein09Extracting,
title = {{Extracting Garch Effects from Asset Returns Using Robust NMF}},
year = "2009",
author = "\href{http://robustandscalable.wordpress.com}{Ruair\'{i} de Fr\'{e}in} and Scott Rickard and Konstantinos Drakakis",
booktitle = "Digital Signal Processing Workshop and 5th IEEE Signal Processing Education Workshop, 2009. DSP/SPE 2009. IEEE 13th",
pages = "200--5",
organization = "IEEE",
doi = "10.1109/DSP.2009.4785921",
urllink = "http://ieeexplore.ieee.org/xpl/login.jsp?tp=&arnumber=4785921&url=http%3A%2F%2Fieeexplore.ieee.org%2Fxpls%2Fabs_all.jsp%3Farnumber%3D4785921",
keywords = "Gaussian noise;learning (artificial intelligence);matrix decomposition;pattern clustering;stock markets;time series;Dow Jones Industrial Average;GARCH effects;asset groupings;asset identification;asset returns;background Gaussian noise;cluster assets;dictionary learning approach;diversified portfolio;factor analysis;financial time series;heteroscedastic latent factor model;non-negative matrix factorization;robust NMF;stock market;univariate ARCH model;volatility clustering characteristics;Adaptive systems;Autocorrelation;Educational institutions;Laboratories;Matrix decomposition;Noise robustness;Portfolios;Probability distribution;Stock markets;Time series analysis;Autoregressive Conditional Heteroscedasticity;Clustering;Low rank decomposition;Non-negative Matrix Factorization;Sparseness",
month = "Jan."}
@INPROCEEDINGS{deFrein14Learning,
author={\href{http://robustandscalable.wordpress.com}{Ruair\'{i} {de Fr\'{e}in}}},
booktitle={IEEE International Workshop on Machine Learning for Signal Processing},
title={{Learning and Storing the Parts of Objects: IMF}},
year={2014},
month={Sep.},
pages={},
keywords={Approximation methods;Convergence;Digital signal processing;Matrix decomposition;Optimization;Quantization (signal);Signal processing algorithms;low rank;nmf;quantization},
doi={10.1109/ICDSP.2014.6900690},
}
@INPROCEEDINGS{deFrein14Quantized,
author={\href{http://robustandscalable.wordpress.com}{Ruair\'{i} {de Fr\'{e}in}}},
booktitle={Digital Signal Processing (DSP), 2014 19th International Conference on},
title={{Quantized Nonnegative Matrix Factorization}},
year={2014},
month={Aug.},
pages={377--82},
keywords={Approximation methods;Convergence;Digital signal processing;Matrix decomposition;Optimization;Quantization (signal);Signal processing algorithms;low rank;nmf;quantization},
doi={10.1109/ICDSP.2014.6900690},
}
@article{deFrein14Integration,
year={2014},
issn={1064-7570},
journal={Journal of Network and Systems Management},
doi={10.1007/s10922-014-9313-9},
title={{Integration of QoS Metrics, Rules and Semantic Uplift for Advanced IPTV Monitoring}},
url={http://dx.doi.org/10.1007/s10922-014-9313-9},
publisher={Springer US},
keywords={Network and system monitoring; Measures; Quality of Service; IPTV; Monitoring; Semantic uplift},
author={\href{http://robustandscalable.wordpress.com}{Ruair\'{i} {de Fr\'{e}in}} and Olariu, Cristian and Song, Yuqian and Brennan, Rob and McDonagh, Patrick and Hava, Adriana and Thorpe, Christina and Murphy, John and Murphy, Liam and French, Paul},
pages={1-36},
language={English}
}
@incollection{deFrein14Multilayer,
year={2014},
isbn={978-3-319-07247-0},
booktitle={Formal Concept Analysis},
volume={8478},
series={Lecture Notes in Computer Science},
editor={Glodeanu, Cynthia Vera and Kaytoue, Mehdi and Sacarea, Christian},
doi={10.1007/978-3-319-07248-7_18},
title={{Multilayered, Blocked Formal Concept Analyses for Adaptive Image Compression}},
url={http://dx.doi.org/10.1007/978-3-319-07248-7_18},
publisher={Springer International Publishing},
author={ \href{http://robustandscalable.wordpress.com}{Ruair\'{i} {de Fr\'{e}in}} },
pages={251--67},
language={English}
}
Significant effort has been placed on analyzing and developing biologically inspired approaches and applying them to computer networking. This tutorial provides a comprehensive overview of biologically inspired approaches for the computer networking field. It emphasizes their primary characteristics as building blocks for designing distributed systems. First we characterize computer networking, its requirements, future perspectives and challenges. Next, we emphasize existing constraints in current networking algorithms and we highlight the advantages of bio-inspired approaches. Biologically inspired approaches for wide to nano computer networking are presented and categorized into “distributed system building blocks”. These building blocks include: communication, coordination and agreement, adaptation, robustness to attacks and failures, and synchronization. For each category/approach, we present the source of biological inspiration, the corresponding mathematical model, primary applications, advantages, limitations and potential directions for future applications. We conclude the tutorial with an overview of potential future directions for bio-inspired algorithms.
@misc {deFrein14Biologically,
title = {{Biologically Inspired Networking Approaches as Building Blocks for Distributed Systems.}},
year = "2014",
author = "Michele Nogueira and \href{http://robustandscalable.wordpress.com}{Ruair\'{i} de Fr\'{e}in} and Tonguz and Ozan",
howpublished = "The 44th Annual IEEE/IFIP International Conference on Dependable Systems and Networks (DSN 2014), Georgia Tech, Atlanta, Georgia USA.",
urllink = "http://2014.dsn.org/tutorials.shtml",
keywords = "Activator-inhibitor systems; Adaptability; Agreement; Artificial immune systems; Attacks and failures; Bio-inspired algorithms; Biological inspiration; Cellular signaling networks; Communication; Complexity; Coordination; Datataxis; Distributed system building blocks; Distributed Systems; Dynamic changes; Epidemic spreading; Failures; Faults; Firefly oscillators; Honey bee dance; Mathematical Model; Networking; Outages; Physiological networks; Quorum Sensing; Robust; Robustness; Scalability; Social insect routing; Synchronization."}
@inproceedings {deFrein13Ghostbusters,
title = {{Ghostbusters: A parts-based NMF algorithm}},
year = "2013",
author = "\href{http://robustandscalable.wordpress.com}{Ruair\'{i} de Fr\'{e}in}",
booktitle = "Signals and Systems Conference (ISSC 2013), 24th IET Irish",
pages = "1-8",
organization = "The Institution of Engineering and Technology - The IET",
doi = "10.1049/ic.2013.0050",
url = "http://repository.wit.ie/2745/1/RdFISSC.pdf",
urllink = "http://repository.wit.ie/2745/1/RdFISSC.pdf",
keywords = "image coding;matrix decomposition;Galois lattice;Ghostbusters;arbitrary lectic ordering;closure operator;column entry;dictionary representation;encoding part-based constraints;nonlinear approximation;nonnegative hierarchical closure structure;nonnegative matrix decomposition algorithm;part-based NMF algorithm;row entry;sparse real-valued dataset;tolerance-to-error constraint;Lectic Orderings;Nonnegative Matrix Factorization;Unique Solutions",
month = "Jun."}
title = {{METHOD AND APPARATUS FOR DETERMINING WHETHER A NODE CAN REPRESENT OR BE REPRESENTED BY OTHER NODES WITHIN A NETWORK}},
year = "2013",
author = "\href{http://robustandscalable.wordpress.com}{Ruair\'{i} de Fr\'{e}in}",
howpublished = "Patent",
url = "http://patentscope.wipo.int/search/en/WO2013007521",
urllink = "http://patentscope.wipo.int/search/en/WO2013007521",
note = "WO Patent 2,013,007,521, Patent Number 2013007521",
month = "18 Jan."}
@incollection {deFrein2013Formal,
title = {{Formal Concept Analysis via Atomic Priming}},
year = "2013",
author = "\href{http://robustandscalable.wordpress.com}{Ruair\'{i} de Fr\'{e}in}",
booktitle = "Formal Concept Analysis",
editor = "Peggy Cellier and Felix Distel and Bernhard Ganter",
pages = "92-108",
organization = "Lecture Notes in Computer Science",
publisher = "Springer Berlin Heidelberg",
doi = {10.1007/978-3-642-38317-5_6},
urllink = "http://dx.doi.org/10.1007/978-3-642-38317-5\_6",
keywords = "Formal Concept Analysis; Rank Reduction; Factorization",
volume = "7880",
howpublished = "http://dx.doi.org/10.1007/978-3-642-38317-5\_6"}
@incollection {Xu2012Distributed,
title = {{Distributed Formal Concept Analysis Algorithms Based on an Iterative MapReduce Framework}},
year = "2012",
author = "Biao Xu and \href{http://robustandscalable.wordpress.com}{Ruair\'{i} de Fr\'{e}in} and Eric Robson and M\'{i}che\'{a}l \'{O} Foghl\'{u}",
booktitle = "Formal Concept Analysis",
editor = "Florent Domenach and DmitryI Ignatov and Jonas Poelmans",
pages = "292-308",
organization = "Lecture Notes in Computer Science, 7278, ",
publisher = "Springer Berlin Heidelberg",
doi = "10.1007/978-3-642-29892-9_26",
urllink = "http://dx.doi.org/10.1007/978-3-642-29892-9_26",
keywords = "Formal Concept Analysis; Distributed Mining; MapReduce",
volume = "7278",
howpublished = "http://dx.doi.org/10.1007/978-3-642-29892-9\_26"}
Rather than using redundancy checking, we keep as many closures as possible in each iteration; All closures are maintained and used to generate the next batch of closures. MRGanter+ has a Map method which calculates local concepts by working on previous concept and local data partition. The Reduce method in MRGanter+ merges local closures first, and then recursively examines if they already exist in the set of global formal concepts H. The set H is used to fast index and search a specified closure; it is designed as a two-level hash table to reduce its costs. The first level is indexed by the head attribute of the closure, while the second level is indexed by the length of the closure.
For more details about MRGanter+, please see our recent publication:
Distributed Formal Concept Analysis Algorithms Based on an Iterative MapReduce Framework:
@incollection{,
year={2012},
isbn={978-3-642-29891-2},
booktitle={Formal Concept Analysis},
volume={7278},
series={Lecture Notes in Computer Science},
editor={Domenach, Florent and Ignatov, DmitryI. and Poelmans, Jonas},
doi={10.1007/978-3-642-29892-9_26},
title={Distributed Formal Concept Analysis Algorithms Based on an Iterative MapReduce Framework},
url={http://dx.doi.org/10.1007/978-3-642-29892-9_26},
publisher={Springer Berlin Heidelberg},
keywords={Formal Concept Analysis; Distributed Mining; MapReduce},
author={Xu, Biao and Fréin, Ruairí and Robson, Eric and Ó Foghlú, Mícheál},
pages={292-308}
}
at http://www.springerlink.com/content/02p8282703rx0m78/. And you can download the author's self-archiving version of this publication from http://arxiv.org/abs/1210.2401.
To know how MapReduce works please refer to "MapReduce Simplified Data Processing on Large Clusters" at http://static.usenix.org/event/osdi04/tech/full_papers/dean/dean.pdf.
Or you can find a presentaion introducing MRGanter+ which we made on ICFCA 2012: http://www.econ.kuleuven.be/ICFCA/present/Biao_Xu.pdf
@ARTICLE{deFrein11Synchronized,
author={\href{http://robustandscalable.wordpress.com}{Ruair\'{i} {de Fr\'{e}in}} and Rickard, Scott T.},
journal={Signal Processing, IEEE Transactions on},
title={{The Synchronized Short-Time-Fourier-Transform: Properties and Definitions for Multichannel Source Separation}},
year={2011},
volume={59},
number={1},
pages={91--103},
keywords={Fourier transforms;parameter estimation;source separation;synchronisation;anechoic mixtures;delay invariant relative windowed-disjoint orthogonality;multichannel source separation;parameter estimation;short time Fourier transform;synchronized linear transform;time frequency domain;time frequency linear transform;Delay;Fourier transforms;Signal processing algorithms;Speech;Synchronization;Time frequency analysis;Signal analysis;source separation},
doi={10.1109/TSP.2010.2088392},
ISSN={1053-587X},}
title = {{Adapting Bases Using the Synchronized Short Time Fourier Transform and Non-negative Matrix Factorization}},
year = "2010",
author = "\href{http://robustandscalable.wordpress.com}{Ruair\'{i} de Fr\'{e}in}",
school = "University College Dublin"}
@inproceedings {deFrein09Learning,
title = {{Learning Speech Features in the Presence of Noise: Sparse Convolutive Robust Non-negative Matrix Factorisation}},
year = "2009",
author = "\href{http://robustandscalable.wordpress.com}{Ruair\'{i} de Fr\'{e}in} and Scott T. Rickard",
booktitle = "Digital Signal Processing, 2009 16th International Conference on",
pages = "1--6",
organization = "IEEE",
doi = "10.1109/ICDSP.2009.5201068",
urllink = "http://ieeexplore.ieee.org/xpl/login.jsp?tp=&arnumber=5201068&url=http%3A%2F%2Fieeexplore.ieee.org%2Fxpls%2Fabs_all.jsp%3Farnumber%3D5201068",
keywords = "matrix algebra;speech enhancement;alpha divergence objective;nonstationary noise;sparse convolutive robust non-negative matrix factorization;speech enhancement techniques;speech features;Additive noise;Background noise;Matrix decomposition;Noise reduction;Noise robustness;Sparse matrices;Spectrogram;Speech analysis;Speech enhancement;Working environment noise;Spectral factorization;Speech enhancement",
month = "Jul."}
@inproceedings {deFrein09Constructing,
title = {{Constructing Time-Frequency Dictionaries for Source Separation via Time-Frequency Masking and Source Localisation}},
year = "2009",
author = "\href{http://robustandscalable.wordpress.com}{Ruair\'{i} de Fr\'{e}in} and Scott T. Rickard and Barak A. Pearlmutter",
booktitle = "Independent Component Analysis and Signal Separation",
editor = " Tulay Adali and Christian Jutten and Joao Marcos Romano and Travassos and Allan Kardec Barros",
pages = "573--80",
organization = "Lecture Notes in Computer Science, 7278,",
publisher = "Springer Berlin Heidelberg",
doi = "10.1007/978-3-642-00599-2_72",
url = "http://www.google.fr/url?sa=t&rct=j&q=&esrc=s&source=web&cd=2&ved=0CC8QFjAB&url=http%3A%2F%2Fwww.bcl.hamilton.ie%2F~barak%2Fpapers%2FICA-2009-constructing.pdf&ei=pZ1LVLasMomrOpTKgZgO&usg=AFQjCNHK8kTyiwRgq0jWwCU62yqyC0uyjg&bvm=bv.77880786,d.ZWU",
urllink = "http://dx.doi.org/10.1007/978-3-642-00599-2_72",
volume = "5441",
howpublished = "http://dx.doi.org/10.1007/978-3-642-00599-2_72"}
@inproceedings {deFrein09Extracting,
title = {{Extracting Garch Effects from Asset Returns Using Robust NMF}},
year = "2009",
author = "\href{http://robustandscalable.wordpress.com}{Ruair\'{i} de Fr\'{e}in} and Scott Rickard and Konstantinos Drakakis",
booktitle = "Digital Signal Processing Workshop and 5th IEEE Signal Processing Education Workshop, 2009. DSP/SPE 2009. IEEE 13th",
pages = "200--5",
organization = "IEEE",
doi = "10.1109/DSP.2009.4785921",
urllink = "http://ieeexplore.ieee.org/xpl/login.jsp?tp=&arnumber=4785921&url=http%3A%2F%2Fieeexplore.ieee.org%2Fxpls%2Fabs_all.jsp%3Farnumber%3D4785921",
keywords = "Gaussian noise;learning (artificial intelligence);matrix decomposition;pattern clustering;stock markets;time series;Dow Jones Industrial Average;GARCH effects;asset groupings;asset identification;asset returns;background Gaussian noise;cluster assets;dictionary learning approach;diversified portfolio;factor analysis;financial time series;heteroscedastic latent factor model;non-negative matrix factorization;robust NMF;stock market;univariate ARCH model;volatility clustering characteristics;Adaptive systems;Autocorrelation;Educational institutions;Laboratories;Matrix decomposition;Noise robustness;Portfolios;Probability distribution;Stock markets;Time series analysis;Autoregressive Conditional Heteroscedasticity;Clustering;Low rank decomposition;Non-negative Matrix Factorization;Sparseness",
month = "Jan."}